The National Acute Brain Injury Study: Hypothermia (NABIS:H) has been examining the effect of moderate hypothermia (33 ) on (Glasgow Coma Score 3-8) since October, 1994. This application requests to extend this trial for 3 years and to expand the sample size from the original 500 patients to 700 patients. This is an important initiative, as no treatment to arrest the secondary biochemical cascade after head injury has yet proven to affect head injury outcome. Recent trials of free- radical scavengers and anti-oxidants have not been successful in demonstrating a significant effect on outcome. As laboratory evidence for these agents was strong, the question of a Type II error must be considered for these trials, and the possibility of this type of error must be anticipated by other investigators. The specific aim of NABIS:H is to determine efficacy and safety of 48 hours of systemic hypothermia, induced by surface cooling, begun within six hours of injury. The study design is a multicenter, randomized, prospective clinical trial of standard management at normothermia vs. standard management with hypothermia. Randomization is stratified at each center by admission GCS. All acute care personnel are blinded to all outcome data. All outcome personnel are blinded to the patient's treatment arm. To date, no outcome evaluations have been done by unblinded personnel. Center data is monitored monthly to detect deviations from the protocol, and blinded A/B data are reviewed by the Performance Monitoring and Safety Board regularly to ensure that the trial is proceeding safety. Patients in NABIS:H must have non-penetrating brain injury, GCS 3-8, motor score on the GCS of 5 or less, and age of 16-65. Also, the time of injury must be known and the patient must be available for cooling within 6 hours. Exclusion criteria include severe multi-trauma, unstable cardiovascular or respiratory status after resuscitation, a GCS of 7-8 with normal CT scan, a GCS of 3 with bilaterally unreactive pupils, and a high probability that the patient will not be available for follow-up. We are currently using waiver of consent at all study centers, except Sacramento. All efforts are made to obtain consent before waiver is used. Patients are managed with a standard protocol of sedation, aggressive management of intracranial pressure/cerebral perfusion pressure, nutrition, and measures to prevent complications. The hypothermia patients will also have skeletal muscle paralysis during hypothermia, and rewarming (about 60 hours). NABIS:H currently has five well-established study centers, and two centers in start up. All centers are part of emergency medical service systems with rapid transport and the resources to achieve protocol compliance. Our current center performance monitoring data indicate that all centers are performing well, with excellent protocol compliance and aggressive enrollment of appropriate patients. Based on past performance, we project a yearly accrual of 150-160 patients/year. We believe that increasing the sample size to 700 patients is necessary to avoid a possible Type 11 error. First, in our initial year, the mean time for patients to reach 33 was 9.5 hours. Second, a recently published Phase II report of moderate hypothermia suggest that this treatment may only be effective in GCS 5-8 patients, and about 20 percent of our current patient group falls in this category. A third factor for increasing the sample size is the disappointing results from other multicenter head injury trials.